Analogues of the Duocarmycins Containing Sulfonate Leaving Groups

Feb 16, 2012 - Ralph J. Stevenson,* William A. Denny, Moana Tercel, Frederik B. Pruijn, and Amir Ashoorzadeh. Auckland Cancer Society Research Centre,...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/jmc

Nitro seco Analogues of the Duocarmycins Containing Sulfonate Leaving Groups as Hypoxia-Activated Prodrugs for Cancer Therapy Ralph J. Stevenson,* William A. Denny, Moana Tercel, Frederik B. Pruijn, and Amir Ashoorzadeh Auckland Cancer Society Research Centre, Faculty of Medical and Health Sciences, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand S Supporting Information *

ABSTRACT: The synthesis of 19 (5-nitro-2,3-dihydro-1Hbenzo[e]indol-1-yl)methyl sulfonate prodrugs containing sulfonate leaving groups and 7-substituted electron-withdrawing groups is reported. These were designed to undergo hypoxia-selective metabolism to form potent DNA minor groove-alkylating agents. Analogues 17 and 24, containing the benzyl sulfonate leaving group and a neutral DNA minor groove-binding side chain, displayed hypoxic cytotoxicity ratios (HCRs) of >1000 in HT29 human cancer cells in vitro in an antiproliferative assay. Four analogues maintained large HCRs across a panel of eight human cancer cell lines. In a clonogenic assay, 19 showed an HCR of 4090 in HT29 cells. Ten soluble phosphate preprodrugs were also prepared and evaluated in vivo, alone and in combination with radiation in SiHa human tumor xenografts at a nontoxic dose. Compounds 34 and 39 displayed hypoxic log10 cell kills (LCKs) of 1.78 and 2.71, respectively, equivalent or superior activity to previously reported chloride or bromide analogues, thus showing outstanding promise as hypoxia-activated prodrugs.



INTRODUCTION The DNA minor groove (adenine N3) alkylating agents exemplified by the natural antitumor antibiotic duocarmycin SA1 (1) and by a range of simpler synthetic precursor seco-form analogues [e.g., hydroxyCBI (seco-1,2,9,9atetrahydrocyclopropa[c]benz[e]indol-4-one) (2, Figure 1)]2 are extremely potent (sub-nM IC50) cytotoxins. Their high potency is considered to be due to their ability to become much more reactive alkylators following initial noncovalent DNA binding3 and to then form DNA monoadducts following a single alkylation event, with these lesions generating little DNA distortion and thus evading some routes of DNA repair. Some analogues of this class [e.g., adozelesin (3) and carzelesin (4), Figure 1]4,5 did undergo clinical trials for cancer treatment but were too myelotoxic. This class has become a major focus for the development of selective prodrug forms, including those designed to release these potent “effectors” by exploiting the specific physiological property of hypoxia in solid tumors. A recent review outlines the challenges, opportunities, and strategies in targeting tumor hypoxia in this regard.6 A CBI analogue (5, Figure 1) with a quinone “trigger” showed no oxic/hypoxic differential, which was suggested to be due to prodrug instability.7 Other analogues (e.g., 6, Figure 1) were tested against cells overexpressing or lacking the two-electron reductase DTdiaphorase, but selective hypoxic cytotoxicity was not evaluated.8 We have previously shown9,10 that nitro analogues (e.g., 7, Figure 1) of duocarmycins are hypoxia-selective, undergoing oxygen-reversible one-electron reduction that ultimately results in generation of the corresponding amino compounds (e.g., 7a, © 2012 American Chemical Society

Figure 1) under hypoxia. The latter amines are potent cytotoxins11 with a similar mechanism to the phenolic secoduocarmycins, alkylating at N3 of adenine in AT-rich regions of DNA,9,12 presumably via an imino form (Figure S1 in the Supporting Information) of the spirocyclohexadienone. In contrast, the hypoxia-activated prodrug (HAP) nitro compounds are relatively nontoxic, consistent with their inability to undergo spirocyclization. The hypoxic cytotoxicity ratios [HCR = IC50(oxic)/IC50(hypoxic)] of compounds such as 7 were variable among cell lines, suggested due to their relatively low one-electron reduction potentials [E(1) = −512 mV for 8, Figure 1].13 Analogues bearing electron-withdrawing groups (EWGs) on ring A (as defined in Figure 1) at the 7-position had the greatest effect [E(1) values from −350 to −420 mV] in raising the reduction potential. The 7-sulfonamide 9 (Figure 1) had HCRs of 19−330 across an 11-cell line panel13 and also presented the opportunity of further modification to provide increased aqueous solubility. The sulfonamide alcohol 10 (Figure 1) showed HCRs from 10 to 250 across a 14-cell line panel, and the corresponding phosphate preprodrug 11 (Figure 1, which rapidly hydrolyzed to 10 in plasma) showed significant hypoxic log10 cell kills [additional cell kill, on a log scale, for the combination of radiation and nitroCBI as compared to radiation alone; log10 cell kills (LCKs) of 0.41−1.82] in mice in combination with radiation (to sterilize the oxygenated cells) in excision assays with five different human tumor xenograft models.14 In SiHa cervical carcinoma xenografts, the combination of 11 and radiation gave complete tumor sterilization in 3/ Received: December 20, 2011 Published: February 16, 2012 2780

dx.doi.org/10.1021/jm201717y | J. Med. Chem. 2012, 55, 2780−2802

Journal of Medicinal Chemistry

Article

Figure 1. (+)-Duocarmycin SA, adozelesin, carzelesin, and some previously reported analogues of the duocarmycins.

shown in Schemes 11−14. The syntheses of precursors and important intermediates are shown in Schemes 1−4. We previously demonstrated that nitroCBIs containing a chloride leaving group and A ring EWGs raise E(1) relative to unsubstituted parent compounds, thus providing HAPs with enhanced reductive metabolism and improved hypoxia-selective cytotoxicity.9,13 While all A ring substituents (i.e., 6-, 7-, 8-, and 9-) raised E(1), the effect was strongest when the substituent was in the 7-position [ΔE(1) = 83−159 mV]. Four examples of 7-substituents (NO2, SO2Me, CN, and SO2NH2) raised E(1) values to above −400 mV. Although higher E(1) alone was not sufficient to observe hypoxic selectivity and significant HCRs (>10) were found for several nitroCBIs (e.g., 7-CONH2) with E(1) < −400 mV,13 the combination of 7-substituents with Hbond donor capacity and higher E(1) generally provided the most highly selective compounds. In the current work, we chose to limit our targets to 7-substituted nitroCBIs. This provides a direct comparison of the new nitroCBIs as HAPs with the 7-substituted prodrugs13,14 and 7-substituted phosphate preprodrugs14 previously reported. It was demonstrated in the chloride series that while the final products are structurally more complex than the corresponding unsubstituted analogues, their syntheses were generally easier.13 This was a direct consequence of the A ring EWGs, which direct nitration of the benzindoline intermediates to the desired 5regioisomer. Although this key step was not completely selective, the major isomer was readily purified by recrystallization or chromatography, and we postulated that this may also be the case in the sulfonate series. In the current study, we

5 mice, whereas the corresponding aminoCBI 10 was completely inactive. Similar trends were observed for 7carboxamide analogues.14 We recently showed that related nitroCBIs bearing a bromide leaving group also provide high HCRs (10−286) in HT29 colon carcinoma and SiHa cell lines, with their phosphate preprodrugs also showing high hypoxic LCKs (0.87−2.80) in SiHa human tumor xenografts.15 Finally,16 we have also compared chloride versus sulfonate leaving groups (e.g., 12 and 13, Figure 1), since the latter are better leaving groups than chloride, with significantly greater nucleofugalities.17 In nitroCBIs lacking an A ring substituent, sulfonates with neutral DNA minor groove-binding side chains18 had consistently higher HCRs (IC50 ratios) than the corresponding chloride compounds; a comparison of 7, 12, and 13 in SKOV3 ovarian carcinoma cells showed HCRs of 2.8, 39, and 246, respectively.16 The general conclusion was that sulfonate leaving groups offer a way of increasing the hypoxic selectivity of nitroCBIs without introducing A ring substituents. In this manuscript, we now systematically explore combinations of sulfonate leaving groups, A ring E(1)-raising 7-EWGs, and various DNA minor groove-binding side chains and evaluate the products as HAPs in vitro and phosphate preprodrugs in vivo.



RESULTS AND DISCUSSION Chemistry. The new nitroCBI prodrugs 14−32 described in this manuscript are listed in Table 1, and their syntheses are shown in Schemes 5−10. New nitroCBI phosphate preprodrugs 33−42 are listed in Table 3, and their syntheses are 2781

dx.doi.org/10.1021/jm201717y | J. Med. Chem. 2012, 55, 2780−2802

Journal of Medicinal Chemistry

Article

Table 1. Cytotoxicity Data for NitroCBIs 14−32 (Sulfonates) and Known14 Chlorides 10 and 144−146a

IC50b (μM) HT29 compd

Z

R

Y

oxic

14 15 16 17 18 19 20 21 22 23 144e 145e 24 25 26 27 28 29 30 31 32 146e 10e

CN CN CN CONH2 CONH2 CONH2 CONH2 CONH2 CONH(CH2)2OH CONH(CH2)2OH CONH(CH2)2OH CONH(CH2)2OH SO2NH2 SO2NH2 SO2NH(CH2)2OH SO2NH(CH2)2OH SO2NH(CH2)2OH SO2NH(CH2)2OH SO2NH(CH2)2OH SO2NH(CH2)2OH SO2NH(CH2)2OH SO2NH(CH2)2OH SO2NH(CH2)2OH

A B C A D E B C A D A D A D A D E A D E F A D

OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph Cl Cl OSO2CH2Ph OSO2CH2Ph OSO2Me OSO2Me OSO2Me OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph Cl Cl

2.6 ± 0.6 2.5 ± 0.3 3.4 ± 0.3 9.2 ± 2.4 2.2 ± 0.1 7.0 ± 0.5 22 ± 7 4.6 ± 1.7 15 ± 3 4.3 ± 0.2 5.7 ± 1.8 12 ± 2 35 ± 4 1.6 ± 0.3 14 ± 1 13 ± 1 26 ± 3 8.4 ± 0.7 38 ± 8 47d 50 ± 2 28 ± 2 9.3 ± 1

SiHa HCRc

hypoxic 0.043 ± 0.01 0.49 ± 0.06 0.62 ± 0.03 0.016 ± 0.006 0.038 ± 0.01 0.17 ± 0.02 0.22 ± 0.05 0.43 ± 0.22 0.61 ± 0.22 0.57 ± 0.19 3.2 ± 0.2 0.11 ± 0.07 0.042 ± 0.007 0.090 ± 0.029 14 ± 7 2.2 ± 0.7 4.1 ± 2.3 0.36 ± 0.03 3.9 ± 0.4 0.95 ± 0.39 >20d 11 ± 2 0.09 ± 0.02

94 ± 24 5.2 ± 1.2 5.4 ± 0.3 1370 ± 502 58 ± 4 40 ± 4 105d 21 ± 14 13 ± 2 12 ± 6 1.7 ± 0.5 485 ± 196 1010 ± 188 18 ± 3 8.6 ± 7.9 7.4 ± 1.7 31 ± 13 23 ± 2 9.7 ± 3.0f 50f 114e 0.67e

HCT116d 13 ± 5 403 ± 280 127 ± 37 193 ± 70

HCT8d

PC3d

55 ± 40 283 ± 201 374e ND

ND ND ND 278 ± 72

a

ND, not determined. bHCR = IC50(oxic)/IC50(hypoxic). Values are means of intraexperiment ratios (±SEM for 2−5 experiments). cSee Table 1. SKOV3, ovarian carcinoma; A549, nonsmall-cell lung carcinoma; C33A, cervical carcinoma; H1299, lung carcinoma; H460, large-cell lung carcinoma; HCT116, colorectal carcinoma; HCT8, colon adenocarcinoma; and PC3, prostate carcinoma. eSingle determination.

d

chose EWGs [Z in Table 1, i.e., CN,13 CONH2,13 SO2NH2,13 CONH(CH2)2OH,14 and SO2NH(CH2)2OH14] based on our established syntheses for chloride analogues and/or improved hypoxia-selective cytotoxicity for chloride analogues.9,13,14 A series of DNA minor groove-binding side chains (R in Table 1) were selected based on our previous work with chloride,13,14 bromide,15 and sulfonate16 analogues. We have recently noted that alternative leaving groups (i.e., bromide15 or sulfonates16) are especially susceptible to halide ion scrambling and elimination in a basic environment. These potential side reactions are enhanced by having EWGs in rings A and B (as

defined in Figure 1). Thus, once sulfonate leaving groups were incorporated into the nitroCBI scaffolds, extra care was required (e.g., use of dilute base and cold temperatures) to avoid elimination of RSO2OH. Generally, we chose to introduce sulfonate leaving groups late in the synthesis. The intermediate 7-cyano acetate 47 was obtained by several routes (Scheme 1), the best of which was halogen exchange of 43 to give iodide 44 followed by protecting group exchange to give 45 and quantitative conversion to 47. This sequence provided the highest overall yield with no separation issues. Nitration of 47 was conducted using concentrated H2SO4 with 2782

dx.doi.org/10.1021/jm201717y | J. Med. Chem. 2012, 55, 2780−2802

E G A D D A D E A D E F A D

CONH2 CONH2 CONH(CH2)2OP(O)(OH)2 CONH(CH2)2OP(O)(OH)2 CONH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2 SO2NH(CH2)2OP(O)(OH)2

19 33 34 35 147j 36 37 38 39 40 41 42 148j 11j

OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph Cl OSO2Me OSO2Me OSO2Me OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph OSO2CH2Ph Cl Cl

Yb 6.87 6.91 6.71 6.92 6.56 7.15 7.18 7.09 6.53 6.92 6.74 7.08 7.29 6.89

0.20 0.36 0.08 0.11 0.20 0.06 0.13 0.11 0.29 0.21 0.23 0.16 0.31 0.26

± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.18 0.18 0.01 0.01 0.12 0.12 0.12 0.18 0.07 0.07 0.07 0.07 0.10 0.12

± ± ± ± ± ± ± ± ± ± ± ± ± ±

7.15 7.15 7.37 7.37 7.36 7.35 7.35 7.15 7.24 7.24 7.24 7.24 7.62 7.52

nitroCBId

controlc NS NS